Ribes Bowker

ARTICLE IN PRESS Information and Organization xxx (2009) xxx–xxx Contents lists available at ScienceDirect Information and Organization journal homepage: www.elsevier.com/locate/infoandorg Between meaning and machine: Learning to represent the knowledge of communities David Ribes a,*, Geoffrey C. Bowker b a Georgetown University, Communication, Culture and Technology, 3520 Prospect St. NW, Suite 311, Washington, DC 20057, United States b Santa Clara University, Science Technology and Society, 500 El Camino Real, Santa Clara, CA 95053, United States article info abstract Article history: Representing knowledge in codified forms is transformative of Received 4 June 2008 ones orientation to that knowledge. We trace the emergence of a Received in revised form 25 March 2009 routine for knowledge acquisition and its consequences for partic- Accepted 20 April 2009 ipants. Over time, participants in the earth science project GEON, Available online xxxx first learned about ontologies and then learned how to create them. We identify three steps in the routine: understanding the Keywords: Ontology development problematic of interoperability; learning the practice of knowledge Routine acquisition; and engaging the broader community. As participants Semantic interoperability traversed the routine they came to articulate, and then represent, Sociology of knowledge representation the knowledge of their communities. In a process we call reappre- Ethnography hension, traversing the routine also transformed participants’ orientation towards their data, knowledge and community, making them more keenly aware of the informational aspects of their fields. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction Ontologies are an information technology for representing specialized knowledge in order to facil- itate communication across disciplines, share data or enable collaboration. In a nutshell, they describe the sets of entities that make up the world-in-a-computer, and circumscribe the sets of relationships they can have with each other. They are a complex and ambitious technical approach to address the problem of diverse languages, heterogeneous categorizations and varied methods for organizing information. In the wake of ontologies the information of a domain is substantially reorganized, facilitating data exchange and reuse. These are the goals for ontologies. Their development is a practical and organizational achievement, and the topic of this paper. We focus on the practical processes surrounding * Corresponding author. Tel.: +1 202 687 4831. E-mail addresses: [email protected] (D. Ribes), [email protected] (G.C. Bowker). 1471-7727/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.infoandorg.2009.04.001 Please cite this article in press as: Ribes, D., & Bowker, G. C. Between meaning and machine: Learning to represent the knowledge of communities. Information and Organization (2009), doi:10.1016/j.infoandorg.2009.04.001 ARTICLE IN PRESS 2 D. Ribes, G.C. Bowker / Information and Organization xxx (2009) xxx–xxx the design and deployment of ontologies within the infrastructure project GEON (the geosciences network) and trace the emergence of an organizational routine for their production. This routinization mirrored the learning trajectory of participants as they came to understand what is at stake in ontologies, that is, that they were representing the knowledge of their communities. As they traversed the routine, participants’ experience of expert knowledge and of their communities were transformed in a process we call reapprehension: an increased orientation to the informational organization of ones field e.g., data, databases, encoded knowledge and their capacity to ‘‘flow” (or interoperate) across techno- logical, disciplinary and institutional divides. Organizational routines are ‘‘repeated patterns of interdependent actions, performed by multiple actors,” (Feldman, 2000; Feldman & Pentland, 2003). They serve as a resource, a malleable and locally adapted recipe or template for how to go about a task. While a routine must always be practically en- acted, having no existence outside its performance, it also becomes embedded in the configuration of material resources that enable practical work (Jordan & Lynch, 1998). We will see that the activities we trace and call the acquisition routine rendered the complex and uncertain activities of knowledge representation into an outline of ‘‘steps,” reducing the work of reinvention on each new occasion of ontology building. However, traversing the routine also changed the experience participants had of their data archive, knowledge and community. In particular we focus on the emergence of a structured concern for ensuring the representativeness of ontologies: the practical methods for creating repre- sentations which stood in for and were used by a larger knowledge community. Ontologies are intended to serve a community, making accessible data and resources for its members; as such they are a form of infrastructure (Star & Ruhleder, 1994). Participants in GEON quickly realized that they comprised a small subsection of the geoscience community (i.e., scores of participants in a community often cited to be in the thousands). Without the work of making ontologies representative of their domain community – of generating venues for feedback and for participation – their ontologies would be open to contestation, or, more likely, be ignored and remain unused. How- ever, modeling ontologies involved articulating knowledge in ways that appeared alien to that domain community. For ontologies to appear representative, the community itself would have to learn the goals and language of knowledge modeling. The phenomenologically transformative consequences of learning and traversing routines are almost completely undiscussed in the literature. In our case, participants’ orientation to information in their dis- cipline was changed by traversing the routine. We name this reapprehension, and emphasize the practice and material tools that accompany the reworking of, for instance, knowledge in informational terms. Participants came to learn: (i) the purposes and goals of ontology, what we call the problematic of interoperability; (ii) how to articulate their knowledge in forms amenable to formal representation, and (iii) how a broader community’s interests are at stake in this process, and what activities would be necessary to engage and enrol that community in the use of ontologies. In order to do so participants had to rear- ticulate their knowledge in forms amenable to formal modeling, and also encourage their colleagues in the ontologies’ use, maintenance and upkeep. A keener awareness of the informational aspects of their fields changed the orientation of participants to their own data and knowledge; it also entailed redirect- ing more time and resources to their integration and maintenance. Following a discussion of case, method and an outline of knowledge capture we trace each of the three steps of the routine. In Section 6 we return to how, by traversing the routine, ‘‘knowledge” and ‘‘community” took on new meaning, as they were rearticulated in the language of logic and information as predicates and users,1 respectively. 1.1. Case and method GEON, the GEOscience Network, is a cyberinfrastructure project (Atkins, 2003) which sought to produce a repertoire of high-end information technologies for the broader earth sciences: 1 There is a close relationship between acquisition and user studies or requirements analysis. In both cases people are recast as users of future systems, the object of studies that make them known, so as to inform a process of technology design (Mackay, Carne, Benyon-Davies, & Tudhope, 2000; Woolgar, 1991). This topic is analyzed more extensively by Ribes and Finholt (2008) in which the authors explore the simultaneous constitution and knowing of a user community. Please cite this article in press as: Ribes, D., & Bowker, G. C. Between meaning and machine: Learning to represent the knowledge of communities. Information and Organization (2009), doi:10.1016/j.infoandorg.2009.04.001 ARTICLE IN PRESS D. Ribes, G.C. Bowker / Information and Organization xxx (2009) xxx–xxx 3 The ultimate goal of GEON is to establish a new informatics-based paradigm in the geosciences, to provide a holistic understanding of the Earth’s dynamic systems, thereby transforming the science (GEON Proposal: 3). The project was funded by the National Science Foundation (NSF) with the goal of producing an ‘‘umbrella infrastructure” bringing together heterogeneous earth science disciplines. It drew together a wide range of earth and computer science experts representing multiple institutions across the US. To name only a few, from the earth sciences GEON includes paleobotanists, metamorphic petrologists and geophysicists; from computer science: database specialists, grid developers and knowledge representation experts (Ribes & Bowker, 2008). The hope was to build tools specific to earth science research but general enough to support work in the various specialties of that community. Ontologies were part of the GEON’s ‘‘knowledge mediation solution,” facilitating communication and collaboration across disciplinary difference. Between 2002 and 2005 we conducted ethnographic research, attending the meetings, workshops and conferences organized by GEON members. The routine we describe emerged as a driving concept of our research as we iteratively returned to the field

Load more